Share your thoughts, 1 month free Claude Pro on usSee more

Chart Question Answering on ChartQAPro 67% held-out (test)

46.88Factoid Accuracy

COGS

Updated 1mo ago

Evaluation Results

Method	Links
COGS 2025.10		46.88	65.73	51.16	61.85	58.25	52.02
In-Context Q Example 2025.10		46.33	62.94	46.91	61.11	61.72	50.58
GPT-5-nano 2025.10		45.95	63.64	49.4	63.58	49.82	50.74
GPT-4o-mini 2025.10		43.63	66.43	45.48	59.88	45.2	48.32
Claude Haiku 3.5 2025.10		43.44	65.03	39.84	61.79	38.77	46.74
Self-Consistency 2025.10		43.44	61.54	44	59.82	41.76	47.22
InternVL3.5-GPT-OSS 2025.10		43.02	58.74	42.86	58.02	54.48	46.86
Chart-R1 2025.10		42.17	46.85	50.53	61.11	55.55	47.32
Decompositional CoT 2025.10		42.08	65.03	42.57	56.53	45.55	46.36
Qwen2.5-VL-7B (base) 2025.10		42.07	62.59	44.88	60.78	50.72	47.36
Gemini 2.5 Flash-Lite 2025.10		40.42	19.96	48.77	37.43	16.66	38.72
Tree of Thoughts 2025.10		40.01	57.94	41.55	54.13	53.35	44.44
ChartQA-Train 2025.10		38.77	60.14	49.72	61.11	53.12	46.64
Phi-4-14B 2025.10		23.18	34.27	40.93	46.91	36.31	31.61
ChartMoE 2025.10		19.03	35.66	32.97	45.68	27.08	27.28
ChartLLaMA 2025.10		8.11	23.08	18.37	45.06	29.55	17.19